Available online at www.sciencedirect.com
Acta Psychologica 127 (2008) 197–210
www.elsevier.com/locate/actpsy
Assessing the symbolic distance effect in mental images constructed
from verbal descriptions: A study of individual differences in the
mental comparison of distances
Michel Denis
*
Groupe Cognition Humaine, LIMSI-CNRS, Université de Paris-Sud, BP 133, 91403 Orsay, France
Received 7 June 2006; received in revised form 16 April 2007; accepted 7 May 2007
Available online 20 July 2007
Abstract
In two experiments, undergraduates processed a verbal description of a spatial configuration on the periphery of which six landmarks
were located. The participants were then invited to generate visual images of the configuration, and to visualize the distances between
pairs of landmarks. Their task consisted of deciding which of the two specified distances was longer. The results showed that as the magnitude of the differences in distance increased, the frequency of correct responses was higher, and response times were shorter. This pattern of results is characteristic of the symbolic distance effect, which is especially interesting in the present experiment where the images
generated by the participants were constructed after processing a verbal description (rather than reconstructed from previous perceptual
experience). In order to assess the role of visual imagery in the comparison of distances, the performance of participants with the highest
scores on a visuo-spatial test (the Minnesota Paper Form Board) was compared to that of those with the lowest scores. High visuo-spatial
imagers had higher frequencies of correct responses and shorter response times than the low imagers in the distance-comparison task.
They outperformed their counterparts even more clearly on the items where the distance differences were the smallest, suggesting that
visual imagery is especially important for items requiring the most attentive examination of a visual image. These data are interpreted
as reflecting the fact that visual imagery mediates the process of mentally comparing distances, even when learning has been essentially
based on verbal input. These findings support the view that a representation constructed from a verbal description may incorporate metric information about distances, and they offer evidence suggesting that visual images constructed from descriptive texts have genuine
analog properties.
2007 Elsevier B.V. All rights reserved.
PsycINFO Classification: 2300; 2340
Keywords: Symbolic distance effect; Mental imagery; Verbal descriptions; Spatial configurations; Mental comparison of distances; Individual differences
1. Introduction
Among the properties that account for the effectiveness
of images in human cognition, their analog character is
the property that has received the most consideration
(e.g., Denis & Kosslyn, 1999; Finke, 1989; Kosslyn, 1980;
Kosslyn, 1994; Paivio, 1971; Paivio, 1986; Paivio, 1991;
Richardson, 1999). Mental images are depictive representa*
Tel.: +33 1 69 85 80 08; fax: +33 1 69 85 80 88.
E-mail address: [email protected]
0001-6918/$ - see front matter 2007 Elsevier B.V. All rights reserved.
doi:10.1016/j.actpsy.2007.05.006
tions of which internal structure is based on a semantics of
resemblance. Not only do mental images contain information, but also this information is organized in meaningful
patterns, and this organization is mapped onto the organization of information in the physical entities represented by
the images. The structural isomorphism between images
and objects or scenes is attested by the fact that they preserve topological relationships between parts of objects,
and even detailed metric information, such as the relative
distances between these parts. The theory of imagery most
often adopted at the present time is based on the concept
198
M. Denis / Acta Psychologica 127 (2008) 197–210
that images can preserve the Euclidean metrics of the
objects perceived (Kosslyn, 1994). This concept lies at the
core of most theoretical hypotheses regarding the architecture and function of the human cognitive system (e.g.,
Guenther, 1998; Kosslyn & Rosenberg, 2001; Shepard &
Cooper, 1982). It is central to Barsalou’s (1999) perceptual
theory of knowledge, and the empirical work supporting the
hypothesis that human cognition is grounded in perception
(e.g., Zwaan & Radvansky, 1998; Zwaan & Taylor, 2006).
Empirical attempts to provide evidence of the analog
character of mental images have been crucial in the development of knowledge about imagery processes. This was
the objective that Stephen Kosslyn assigned to the imagescanning paradigm (Kosslyn, 1973; Kosslyn, Ball, & Reiser, 1978; Kosslyn, Margolis, Barrett, Goldknopf, & Daly,
1990; Pinker & Kosslyn, 1978). Scanning experiments have
shown that when people are invited to scan mentally across
the visual image of a spatial configuration, the greater the
distance between two points in that configuration, the
longer it takes them to scan the image. This chronometric
consistency is viewed as evidence of the metric qualities
of the mental representation to which the scanning process
is applied, and taken as the empirical signature of the structural isomorphism between visual images and the configurations they represent (cf. Borst, Kosslyn, & Denis, 2006;
Iachini & Giusberti, 2004; Mellet et al., 2000; Pinker, Choate, & Finke, 1984). A further value of the image-scanning
paradigm, in this context, is that it applies not only to mental images of previously perceived objects, but also to the
images constructed from verbal descriptions of configurations, that is, to the images constructed in the absence of
any current or recent perceptual input (Chabanne, Péruch,
Denis, & Thinus-Blanc, 2004; Denis & Cocude, 1989;
Denis & Cocude, 1992; Denis & Cocude, 1997; Denis,
Gonçalves, & Memmi, 1995; Mellet et al., 2002).
However, the image-scanning paradigm has been challenged on theoretical grounds, in particular because it
was suspected of reflecting what people know about the
relationships between time, speed, and distance, rather
than being a genuine property of the images themselves
(Pylyshyn, 1973; Pylyshyn, 2002; Richman, Mitchell, &
Reznick, 1979). The argument of the ‘‘cognitive penetration’’ of images by knowledge and beliefs has been countered by further empirical data (e.g., Denis & Carfantan,
1985; Finke & Pinker, 1982; Kosslyn, Pinker, Smith, &
Shwartz, 1979; Reed, Hock, & Lockhead, 1983). Nevertheless, several problems remain with this paradigm, which
leaves data collected using this method open to criticism.
One of these problems is the lack of consistency in how
people interpret instructions about how to ‘‘scan’’ an
image. Although the instructions are usually carefully compiled by the experimenters, they remain open to diverse
interpretations by the participants, a situation that could
clearly undermine the empirical value of data collected
from participants who have differing perceptions of the
task. Another problem with image-scanning experiments
is that the participants’ responses do not provide any mea-
sure of ‘‘performance’’. The participants are just asked to
‘‘declare’’ that they have completed an image-scanning
trial, and the time that has elapsed when they make such
declaration is taken to be the time they have actually
devoted to scanning. This is a problem if a chronometric
measure is conceived as a genuine reflection of a cognitive
process. Interestingly, however, in spite of these problems,
image-scanning experiments have provided quite consistent
and dependable results, from which legitimate claims about
the process of mental imagery have been generated, but we
still need a behavioral measurement that would make it
possible to assess the analog character of the mental representations constructed in mental imagery in an unambiguous fashion. To achieve maximum validity, such a
measurement should be collected for a task where participants are invited to process a metric property of the imagined object, and which offers the experimenter a way of
comparing their performance with an objective norm.
This condition would be met by a paradigm in which
participants are required to form the visual image of a spatial configuration, and then assess a metric property of the
configuration, for instance, comparing two distances and
deciding which is the longer (or the shorter) of the two.
Instead of requiring participants to perform absolute estimations of distances, the task now consists of comparing
two distances, a task which is very easy to describe in the
instructions, and unlikely to elicit diverse interpretations
from the participants. Moyer (1973) found that when
people are presented with simple physical stimuli, such as
straight lines displayed side by side, the greater the difference between the two lengths, the easier the decision, and
in particular, the shorter the time taken to respond (see
Moyer, 1973; Moyer & Bayer, 1976). This phenomenon
was labeled the ‘‘symbolic distance effect’’, and considered
to reflect what Moyer called an ‘‘internal psychophysics’’.
Subsequent experiments showed that this effect was maintained when participants compared the sizes of familiar
objects from memory (e.g., Dean, Dewhurst, Morris, &
Whittaker, 2005; Marschark, 1983; McGonigle & Chalmers, 1984; Paivio, 1975), or even performed novel, unfamiliar comparison tasks on familiar objects (such as
comparing the angles formed by the two hands of clocks;
see Paivio, 1978; Trojano et al., 2002; Trojano et al.,
2000; Trojano et al., 2004). Such findings indicated that
the consistencies and constraints that govern the processing
of visual stimuli also apply to the corresponding internal
representations. In other words, the internal representations exhibiting such consistencies and constraints meet
the requirements that allow us to classify them as analog
representations.
A further step in generalizing the symbolic distance
effect consisted of searching for evidence collected when
people mentally process novel objects (i.e., for which no
pre-stored information is likely to facilitate mental comparisons), and specifically an object entirely constructed from
verbal information. Denis and Zimmer (1992) reported an
experiment (Expt. 2a) in which participants were invited to
M. Denis / Acta Psychologica 127 (2008) 197–210
learn a spatial configuration either from a visual presentation of a map or from processing a description of it. The
configuration included six geographical details that were
positioned at its periphery at locations that the description
made explicit by using the clock-face conventions of air
navigation (e.g., ‘‘At 11 o’clock, there is a harbor’’). After
learning the map or the description, the participants were
asked to construct an image of the configuration and to
decide which of two designated distances (for instance,
‘‘harbor–lighthouse’’ and ‘‘harbor–beach’’) was longest.
Not surprisingly, a distance effect was obtained after map
learning, so that the mean frequency of correct responses
increased as the magnitude of the difference being assessed
increased. More interesting was the fact that a distance
effect was also found for the participants who had constructed the mental representation from the verbal description. It is important for our argument that correct
judgments about distances were produced (reflecting the
chronometric regularities of the symbolic distance effect)
even when the distances had not been experienced visually
during learning, and had not even been made explicit in the
description. Nevertheless, their metric values appeared to
have been incorporated in the representation that the participants constructed while encoding the description. The
fact that the same pattern was identified not only when
distance judgments were based on physical stimuli, but
also when they were based on their imaged counterparts
(whether these images had been reconstructed from perception or constructed from a description), attests to the analog nature of visual images.
This finding was thought to be of special interest since
this was the first time that distance effects had been
obtained in a situation where the representation was not
retrieved from long-term memory, but resulted from episodic construction. This was a clear indication that a representation constructed from a verbal description may
incorporate metric information about distances, and
offered evidence in favor of the analog properties of visual
images constructed from descriptive texts. In a further
experiment (Denis & Zimmer, 1992, Expt. 2b), which was
restricted to the condition based on the learning of a verbal
description, chronometric measures corroborated the previous findings. The time taken to produce correct responses
decreased as the difference between the distances being
compared increased. Further support was provided by
experiments using variants of the distance-comparison paradigm (Afonso, Gaunet, & Denis, 2004; Péruch, Chabanne,
Nesa, Thinus-Blanc, & Denis, 2006).
Recently, Noordzij and Postma (2005) broadened the
perspective on the symbolic distance effect following learning of spatial descriptions by using more complex verbal
materials than those of Denis and Zimmer (1992). The texts
used in their experiment were spatial descriptions of realistic environments, such as a zoo or a shopping center. When
invited to compare bird-flight distances between pairs of
objects, the participants performed faster and better with
increasing distance differences, which reflected the presence
199
of a symbolic distance effect. An interesting addition to this
experiment was the contrast between two modes of description, namely, survey and route descriptions. Regardless of
the type of description that the participants had listened to,
negative correlation coefficients were found between the
metric distance differences and the comparison times. Finer
analyses indicated that survey descriptions had a relative
advantage, suggesting that they lead to mental spatial representations that pinpoint the location of the objects more
accurately than route descriptions (see also Noordzij, Zuidhoek, & Postma, 2006; Péruch et al., 2006).1
However, the studies reported above left several questions unanswered, which the experiments reported below
were designed to answer. The first question pertains to the
assumption that mental imagery mediates the process of
mentally comparing distances. The involvement of imagery
in the mental comparison of objects is attested by the fact
that performance declined in response to interference resulting from a secondary task making demands on the participants’ visual working memory (cf. Dean et al., 2005).
Evidence from other sources is also required. In particular,
if mental imagery is the mediating process in the mental
comparison of distances (even when the original source of
information is verbal in nature), this should be reflected
by differences in performance depending on individual
imagery capacities. The experiments reported here were
designed to establish whether such individual differences
would predict performance in distance comparison. If high
visuo-spatial imagers were to outperform low imagers, this
would support the interpretation that distance comparisons
are genuinely based on the process of visual imagery, even
when the information that imagery provides has been constructed from verbal discourse. An individual differences
approach was used by Paivio (1978) in a study of the symbolic distance effect, which involved the mental comparison
of the sizes of angles, where high imagers did perform significantly faster than low imagers. No such evidence has been
collected in the domain of distance comparisons, and this
was the primary objective of the present study. Recording
data in a distance-comparison task was worthwhile in that
participants were invited to focus on exactly the same metric property as that considered in image-scanning studies,
i.e., distances between points in a two-dimensional space.
A further objective of the present study reflects our interest in an unexplored facet of the issue under investigation.
1
Reference to the metric properties of mental images is common in the
imagery literature. In the case of distance comparisons making use of
imagery, there is no difficulty in recognizing that people have to use metric
information. However, strictly speaking, the term ‘‘metric’’ implies the
application of some numerical value combined with a unit of measurement. Admittedly, the comparison task can be executed accurately
regardless of the scale of a participant’s representation. It would therefore
be more correct to refer to a notion of ‘‘relative metrics’’. The aspect of an
internal representation endowed with metric properties that is relevant
here is that the values reflecting the various distances are consistent with
one another within the representation, even if there is no objective metric
counterpart to these values in the outer world.
200
M. Denis / Acta Psychologica 127 (2008) 197–210
There is a strong presumption that imagery is involved in
distance comparisons, but it is also likely to be involved
at earlier steps, during the construction of the representation, that is, during the processing of the verbal description.
If this is a valid assumption, participants with the highest
visuo-spatial capacities should not only exhibit the benefits
they enjoy from imagery while they are making the comparisons, but also during the earlier stage when they were constructing the representation. The objective of our second
study was to collect fresh data about the visual imagery processes involved during the learning of spatial descriptions.
Fine chronometric measurements can be expected to be
affected by the participants’ imagery capacities, with high
visuo-spatial imagers displaying faster processing.
2. Experiment 1
Experiment 1 replicated the part of Denis and Zimmer’s
(1992) Experiment 2a involving the processing of a verbal
description. After learning a description, the participants
were invited to compare pairs of distances, and the frequency of their correct responses was calculated. The new
data were expected to confirm an increase in the frequency
of correct responses as the magnitude of the differences
between the distances being compared increased. The new
piece of evidence that this experiment was intended to provide was the superior performance of high visuo-spatial
imagers. If this were to be observed, it would support the
view that imagery effectively mediates the comparison process, since the participants who made the best use of their
imagery capacities would outperform participants who
were less prone to rely on images.
2.1.2. Materials
A text was written describing the map of a circular
island on the periphery of which six geographical landmarks were located. The locations of these landmarks were
defined in the terms of the air navigation clock-face convention. They were introduced in clockwise order, starting
with the landmark situated at the 11 o’clock position. The
six landmarks were located in such a way that the distances
between pairs of adjacent items were all different. The text
read as follows (original in French): ‘‘The island is circular
in shape. Six features are located at its periphery. At 11
o’clock, there is a harbor. At 1 o’clock, there is a lighthouse.
At 2 o’clock, there is a creek. Midway between 2 and 3
o’clock, there is a hut. At 4 o’clock, there is a beach. At 7
o’clock, there is a cave.’’ In French the names of the landmarks were all one-syllable words.
The list of all the distances between pairs of landmarks
was established, including both of the possible formulations for every distance (e.g., ‘‘harbor–beach’’ and
‘‘beach–harbor’’). The materials for the comparison task
involved a subset of these pairs of distances. The pairs
selected were those that used the first landmark in the
two distances to be compared (e.g., ‘‘harbor–beach’’/‘‘harbor–hut’’). Two comparisons involving equal distances
were excluded (‘‘hut–lighthouse’’/‘‘hut–beach’’, and
‘‘beach–cave’’/‘‘beach–lighthouse’’). The 58 items were
listed in random order. The constraints were that the same
distance could not occur more than three times in successive items, and that no more than three items requiring
the same response could occur in a row. A second list
was constructed by reversing the order of the two distances
in each pair (for instance, ‘‘harbor–beach’’/‘‘harbor–hut’’
was replaced by ‘‘harbor–hut’’/‘‘harbor–beach’’).
2.1. Method
2.1.1. Participants
The participants were 20 undergraduates from the Orsay
campus. They were a subset of a larger sample of undergraduates (N = 69) who had been collectively tested using a set of
five visuo-spatial tests and questionnaires one week earlier in
a lecture room of the university. One of the tests was the
Minnesota Paper Form Board (MPFB; Likert & Quasha,
1941), a visuo-spatial test widely used in imagery research
(e.g., Denis & Cocude, 1997; Paivio, 1978; Pazzaglia & De
Beni, 2001). In the present sample, the MPFB scores ranged
from 10 to 30. The mean for the group was 18.3 (sd = 3.8)
and the median was 18.5. From among the 69 participants,
the 10 who had the lowest scores and the 10 who had the
highest scores were invited to take part in the experiment
described below. Care was taken to include equal numbers
of male and female participants in the two groups (hereafter
called MPFB and MPFB+, respectively).2
2
There were equal numbers of male and female participants in this and
the next experiment. In fact, Gender was never significant, and has
therefore been left out of the report of the analyses.
2.1.3. Procedure
2.1.3.1. Learning. The participants were informed that they
would listen to a description of the map of an island, of
which they were asked to create as vivid and accurate a
visual image as possible. They were seated in front of a
board placed 140 cm from them. On this board was a blank
sheet of paper (75 · 75 cm), which was intended to provide
a cue for the imaged size of the island. The board was then
removed. The description was provided in auditory form as
a tape recording played three times. Following each hearing, the participants were required to form a visual image
of the map and to focus mentally on each of the geographical landmarks in turn and check the exact location of each.
2.1.3.2. Distance comparisons. The participants were given
two pages listing pairs of distances written side by side
(in the form of landmark names designating two distances
to be compared). They were instructed to picture the entire
map mentally, and then to visualize the two distances designated and compare them. The participants were then
asked to respond by ticking the pair of landmark names
that corresponded to the longer of the two distances. They
were asked to respond as fast as possible, without impair-
M. Denis / Acta Psychologica 127 (2008) 197–210
201
Capacities as between-participant factors, and Magnitude
of Distance Differences as within-participant factor. Overall performance was significantly lower for MPFB than
for MPFB+ participants, 0.85 (sd = 0.08) vs. 0.91
(sd = 0.04), respectively, F(1, 16) = 5.25, p < .05. A significant effect of the magnitude of distance differences was
found, F(2, 24) = 42.81, p < .001. Tukey’s post hoc tests
showed that there were significantly fewer correct
responses to D1 than to D2 items, and to D2 than to D3
items (in both cases, p < .001). Furthermore, there was a
significant interaction between the two variables,
F(2, 32) = 3.17, p = .05. Inspection of the data revealed
that performance was significantly lower in MPFB than
in MPFB+ participants for the D1 items (Tukey,
p < .005), but that the difference remained below the level
of significance for the D2 and D3 items.
Further analyses were conducted after withdrawing
some items from the data, in which the first distance designated was either the longest or the shortest of all possible
distances. For these items, the response was indeed easy
to predict. For instance, if the first mentioned distance
was the longest of all (‘‘lighthouse–cave’’), a smart participant would not have to visualize the second distance to
decide that it was shorter than the first one. Withdrawing
these items reduced the number of items from 17 to 13 in
D1, from 21 to 17 in D2, and from 20 to 12 in D3. However, after these corrections, the frequencies of correct
responses remained remarkably similar to those computed
in the first set of analyses, and the new set of ANOVAs
simply confirmed the previous ones.
ing the accuracy of their responses. Half of the participants
received the first list of pairs and half received the list of
reversed pairs. At the end of the task, the participants were
interviewed. Two participants who reported having followed the imagery instructions for less than 75% of the
time during the task were excluded and replaced. In addition, the participants were asked whether before mentally
comparing the distances they had either relied on the location of the landmarks depicted in their visual image or first
revised the hour-coded location of the landmarks. None of
the participants reported having used this latter procedure.
3. Results
The items were divided into three subsets, depending on
the magnitude of the difference between the two distances
to be compared. Differences were expressed in terms of
the actual differences on the map used in the Denis and
Zimmer (1992) experiments. The first set of items (D1)
comprised 17 items involving distance differences of less
than 10 cm. The second set (D2) included 21 items involving differences of between 10 cm and 20 cm. The third set
(D3) comprised the remaining 20 items, all of which
involved differences of more than 20 cm. If these values
were re-expressed in terms of ratios to the diameter of
the circular map, the D1 items involved distance differences
comprised between 0.03 (i.e., the smallest distance difference, ‘‘cave–creek’’/‘‘cave–lighthouse’’) and 0.17; the D2
items involved differences comprised between 0.17 and
0.34; and the D3 items involved differences comprised
between 0.34 and 0.83 (i.e., the largest distance difference,
‘‘creek–cave’’/‘‘creek–hut’’).
Fig. 1 shows the mean frequencies of correct responses
in the distance-comparison task for each set of items and
each imagery group. An analysis of variance (ANOVA)
was conducted on the data, with Gender and Imagery
3.1. Discussion
The symbolic distance effect, which Denis and Zimmer
(1992) initially reported to occur when people process
episodic images constructed from descriptions of novel
0.97
1.00
0.98
0.91
Frequency of correct responses
0.84
0.87
0.80
0.72
0.60
MPFBMPFB+
0.40
0.20
0.00
D1
D2
D3
Magnitude of distance differences
Fig. 1. Mean frequency of correct responses as a function of the magnitude of distance differences for MPFB
and MPFB+ participants (Experiment 1).
202
M. Denis / Acta Psychologica 127 (2008) 197–210
objects, was confirmed by the data of the present experiment. The clear hierarchy in the frequency of correct
responses in a distance-comparison task as a function of
increasingly demanding comparisons attests to the analog
character of the representations used by the participants
as the basis of their judgments, and this occurred despite
the fact that no perceptual experience was at the origin
of these representations. The representations constructed
via mental imagery can therefore be conceived of as structural analogs of those that typically remain in the memory
after perceptual episodes.
The second finding of interest is the evidence that individual imagery capacities had a measurable effect on performance in the distance-comparison task, where people
who scored low on a visuo-spatial test were clearly outperformed by those with high imagery scores. This is the first
time such an effect has been demonstrated in the distancecomparison task, which attests to the effectiveness of imagery in comparing distances constructed by imagination.
The higher performance of high visuo-spatial imagers
should not be interpreted as indicating that subjects classified as low imagers do not rely on visual imagery at all
when trying to respond to the distance-comparison test.
The latter group did in fact claim that they used imagery,
as instructed, just as the former group did. What was presumably happening was that they had more difficulty in
accessing and exploiting the analogical information provided by the imagery process. Low imagers were overall
less often correct, which confirmed that they experienced
greater difficulty in a task expected to call upon mental
imagery.
A more subtle effect that merits consideration is the fact
that high visuo-spatial imagers’ superiority was not of the
same magnitude for all items. Their superior performance
was most marked for the most difficult items, those that
involved narrow differences and were thus thought to
require the most attentive scrutiny of the image. Not surprisingly, the participants who had the highest imagery
capacities enjoyed a cognitive advantage, but this advantage was markedly lower for the items of medium difficulty,
and not perceptible at all for the easiest items. This pattern
of results is highly meaningful as it reveals that imagery
capacities are exploited to a greater extent when distance
differences are difficult to estimate. There was no difference
in performance between the high and the low imagers for
the items where the difference between the two distances
being compared was very obvious from the image.
The metric accuracy of images constructed from descriptions, which is closely connected to their analog character,
is a relevant feature in the present experimental context.
High imagers are able to construct representations of
which the geometrical properties are devoid of fuzziness.
In the present task, their greater ability to process visual
images is in line with the previous demonstration that in
image scanning, people with the highest visuo-spatial
capacities are those who display a typical pattern of image
scanning (relatively short scanning times and significant
time/distance correlations). Conversely, participants with
poorer visuo-spatial capacities produce responses with
chronometric characteristics that do not suggest that their
images possess any stable, consistent structural properties
(cf. Denis & Cocude, 1997).
Lastly, the absence of any consequence of withdrawing
the responses to predictable items from the data analysis
is a good indication that the participants did follow the
imagery instructions to base their judgments on images,
and did not attempt to perform computations of some
other type. Had this been the case, the elimination of the
items thought to be insensitive to the difficulty of the comparison would have had an impact on the data. The
absence of any change therefore confirms that the participants did indeed deal with the entire set of comparisons
using the same imagery process.
4. Experiment 2
Experiment 1 revealed distinct patterns of results for the
participants identified as being particularly effective at generating and manipulating visual images, versus those who
were less prone to engage in mental imagery. When both
groups were invited to compare distances within a representation constructed from a verbal description, the former
were shown to perform significantly better than the latter,
but both groups displayed a similar significant symbolic
distance effect. However, in addition to investigating the
frequency of correct responses, we also needed to assess
the distance effect by means of appropriate chronometric
measures. This was the primary objective of Experiment 2.
Another objective of this experiment was to extend the
measures made during the distance-comparison task to
the earlier phase of constructing the images to be used in
the comparison task. A number of experiments involving
on-line chronometric measures have revealed the progressive nature of the construction process, and have confirmed
that individual differences are highly likely to appear during the construction of the representation (Denis &
Cocude, 1992; Denis et al., 1995). In particular, in the present
situation, high imagers might prove to be especially efficient
in constructing images, just as they are more efficient in
manipulating them later (during the comparison task). The
idea was therefore to record processing times during the
acquisition of the description. To do this, the participants
were invited to read the sentences describing the configuration, and their processing times were recorded, with the
expectation that these might reflect the difficulties experienced during encoding, in particular by low imagers. The
aim was to assess whether high imagers would perform better
during encoding, as well as in the distance-comparison task.
Another way to test the differential capacities of participants during the encoding of the description is to expose
them to versions of the descriptions that are more or less
likely to assist them in constructing analog images. Previous experiments on image scanning and memorizing spatial
descriptions have shown that distortions imposed on the
M. Denis / Acta Psychologica 127 (2008) 197–210
structure of descriptions create cognitive difficulties at
encoding and, as a result, impair the structural quality of
the constructed images (Denis & Cocude, 1992; Denis &
Denhière, 1990). One way to make image construction
more difficult consists of delivering the information in a
sequence that does not fit the readers’ expectations, for
instance by presenting pieces of information in a random
order rather than in a systematic one (e.g., proceeding
clockwise). In the present experiment, low imagers were
expected to experience greater difficulty than their high
imager counterparts in processing a poorly organized
description.
To summarize, the present experiment investigated the
performance of high and low imagers during two phases,
namely, during the reading of a spatial description (when
the participants were engaged in the process of constructing a visuo-spatial representation of the described
configuration) and subsequently during the distance-comparison task (when they were retrieving the corresponding
information and making their comparisons). We expected
that during encoding, difficulties would be more evident
for low than for high imagers, and that difficulties would
also be greater when the participants had to process a
poorly structured description rather than a well-organized
one. Later on, these difficulties would be expected to manifest themselves again when the participants were engaged
in the comparison task. Lower frequencies of correct
responses and longer response times for comparisons
involving small distance differences would be expected to
confirm the symbolic distance effect obtained in Experiment 1, and these effects should be more marked for
low than for high imagers, given their special sensitivity
to particularly demanding comparisons involving fine discrimination of lengths.
Finally, in order to monitor the process through the
experimental session, processing times were recorded during successive learning trials, and the comparison task
was performed twice: once after three learning trials, and
once again after three further trials.
4.1. Method
4.1.1. Participants
The participants were 32 undergraduates from the Orsay
campus, half of them male and half female. None of them
had taken part in the previous experiment.
4.1.2. Materials
The text used in Experiment 1 was used again, but it was
now presented in written form. This text will be referred to
below as Text 1. This text was used to construct Text 2, in
which the six geographical details were introduced in
random order. Text 2 read as follows: ‘‘The island is circular in shape. Six features are situated at its periphery. At 11
o’clock, there is a harbor. At 4 o’clock, there is a beach. At 1
o’clock, there is a lighthouse. At 7 o’clock, there is a cave.
203
Midway between 2 and 3 o’clock, there is a hut. At 2 o’clock,
there is a creek.’’
The materials for the comparison task involved a subset
of the pairs of distances used in Experiment 1, namely, 10
D1, 10 D2, and 10 D3 pairs. The 30 pairs were arranged in
random order to constitute List 1, following the same constraints as those used in Experiment1. For half of the pairs
in List 1, the first distance designated was shorter than
the second, whereas for the other half, the second distance
designated was longer. List 2 was constructed by reversing
the order of the two distances in each pair (for instance,
‘‘harbor–cave’’/‘‘harbor–creek’’ in List 1 was replaced by
‘‘harbor–creek’’/‘‘harbor–cave’’ in List 2). Furthermore,
each of the two lists was arranged in two sequences, that
is, from the first to the last item as in the original random
listing (Order A) or in the reverse order, from the last to the
first item (Order B). A tape recording was made to present
the distance-comparison test. For each trial, the first two
names of landmarks designating a given distance were
recorded, followed 2 s later by the two names designating
the other distance that was to be compared to the first
one. Presentation of the last name started a clock. The
whole procedure was driven by a computer program
designed specifically for the purpose of the experiment.
Equal numbers of male and female participants (2
males, 2 females) were allocated to the eight conditions
resulting from the combination of the two versions of the
description (Texts 1 and 2), the two lists of items used during the comparison task (Lists 1 and 2), and the two orders
of presentation of these items (Orders A and B).
4.1.3. Procedure
The participants were invited to read one version of the
description three times in immediate succession. After
reviewing the learned configuration mentally, they performed the distance-comparison task. The whole procedure
was repeated, with three more readings of the description,
followed by a mental review and the distance-comparison
task. The procedure was thus basically the same as in
Experiment 1, except that the description of the island
was presented in written form in the learning phase, and
the materials were presented in auditory form in the distance-comparison task.
4.1.3.1. Learning. At the beginning of the first learning
phase, the participants were told that they were to read a
description of the map of an island. They were told that
they would be required to create as vivid and accurate a
visual image of the map as possible. They were seated in
front of a computer screen on which the description would
be displayed sentence by sentence. They were invited to
press the space bar when they wanted to go on to the next
sentence. They were told that there would be no way of rereading any previous sentence. They were therefore invited
to take as long as they felt necessary in reading each sentence before going on to the next one. The time taken to
process each sentence was recorded, but the participants
204
M. Denis / Acta Psychologica 127 (2008) 197–210
were not aware of this. At the end of the first reading episode (R1), two more readings (R2 and R3) were carried out
under the same conditions. In the second half of the experiment, three further reading episodes (R4–R6) took place
with the same instructions and the same procedure.
4.1.3.2. Mental review. After the third reading of the
description, the participants were asked to review the constructed representation mentally, that is, to situate each
landmark at its specific location in their mental image.
The participants were invited to signal to the experimenter
as soon as they had completed this reviewing process and
were ready to go on to do the distance-comparison test.
The time taken to review the description mentally was
recorded. In the second half of the experiment, the same
mental review of the description took place after the last
reading episode.
4.1.3.3. Distance comparisons. After mentally reviewing the
configuration, the participants were introduced to the comparison task. They were told that each trial would consist
of hearing the names of two landmarks on the map from
a tape. They were to picture the entire map mentally, and
then focus on the straight-line distance between the two
landmarks designated. This distance was to be used as
the standard against which a second distance would be
compared. After an interval of 2 s, the names of two landmarks defining another distance were presented. The participants were to visualize this distance while keeping the
first one clearly in their minds, and had to decide whether
the second distance was longer than the first. Presentation
of the first name of the second distance started a clock. The
participants were provided with two buttons. They were to
press one of these buttons if the answer was ‘‘yes’’ (i.e., if
the second distance was longer than the first one), and
the other button if the answer was ‘‘no’’ (i.e., if the second
distance was shorter). The clock was stopped when either
button was pressed, and response time was recorded. The
participants were asked to respond as quickly as possible,
but without impairing the accuracy of their responses. Half
of them were required to use their dominant hand for positive responses, and the other hand for negative responses.
The opposite assignment of hands to positive and negative
responses was required from the other half of the participants. For each participant, equal numbers of items were
expected to elicit Yes and No responses. The participants’
responses were recorded via the computer. Two participants who reported having followed the imagery instructions for less than 75% of the time during the task were
excluded and replaced. None of the participants said that
they had revised the hour-coded locations of the landmarks
before mentally comparing the distances.
4.1.3.4. Visuo-spatial imagery test. At the end of the experimental session, the participants completed the MPFB. The
scores on this test ranged from 9 to 27. The mean was 21.0
(sd = 4.1) and the median was 21.1.
4.2. Results
4.2.1. Learning
Processing times for each sentence during each of the six
reading episodes were computed. Below, we only consider
the times for sentences stating landmarks’ positions (thus
excluding the two introductory sentences). Any individual
times that were more than 2.5 sd above the mean time
for the corresponding group in a given reading episode
were replaced by that limit value (m + 2.5 sd). Only three
such replacements were carried out (for one participant
in the third reading episode, and for another in the fourth
and the fifth reading episodes). Such cases corresponded to
only 1.6% of the total number of individual processing
times.
An ANOVA was performed on the processing times,
with Gender and Text as between-participant factors, and
Reading Episode as within-participant factor. Text 1 took
less time to read than Text 2, 6.30 s (sd = 1.88) vs. 6.76 s
(sd = 2.00), respectively, but this difference was not significant. A significant effect of the rank of the reading episode
was found, F(5, 140) = 21.36, p < .001. On average, the first
three reading episodes took longer to process than the last
three, 7.89 s (sd = 2.16) vs. 5.17 s (sd = 2.72), respectively,
F(1, 28) = 23.47, p < .001. Tukey’s post hoc tests showed
that processing times were longer for R1 than for any of
the other reading episodes (p < .001), and that the only
other significant differences were those between R2 and
both R5 (p < .02) and R6 (p < .005). There were no significant interactions among the main variables.
The sample of participants was then considered as a
whole set from which two subgroups with contrasting
imagery capacities were composed. To do this, we withdrew from the analysis the participants whose scores fell
in the region of the median of the MPFB scores. We only
retained those participants with MPFB scores equal to or
lower than 20 (MPFB , N = 13, average score: 17.2) and
those with MPFB scores equal to or higher than 23
(MPFB+, N = 12, average score: 24.9). An ANOVA was
conducted with Text and Imagery Capacities as betweenparticipant factors, and Reading Episode as within-participant factor. This analysis did not reveal any significant difference between low and high imagers’ processing times,
6.49 s (sd = 1.66) vs. 6.53 s (sd = 2.01), respectively.
4.2.2. Mental review
The times taken by the participants to mentally review
the configuration were recorded after the third reading episode (Review 1) and the sixth one (Review 2). An ANOVA
of the reviewing times was conducted, with Gender and
Text as between-participant factors, and Review as
within-participant factor. Overall, the representation constructed from Text 1 took less time to review than that constructed from Text 2, 16.84 s (sd = 11.24) vs. 25.18 s
(sd = 13.55), respectively, F(1, 28) = 3.65, p < .07, and
Review 1 took longer than Review 2, 30.28 s (sd = 18.65)
vs. 11.74 s (sd = 13.53), respectively, F(1, 28) = 35.12,
M. Denis / Acta Psychologica 127 (2008) 197–210
p < .001. There was a significant interaction between Text
and Review, F(1, 28) = 6.01, p < .05. Post hoc tests revealed
that reviewing times were significantly shorter for Text 1
than for Text 2 during Review 1 only, 22.28 s (sd =
15.46) vs. 38.28 s (sd = 18.52), respectively (p < .01), but
that they did not differ from each other during Review 2.
The next analysis contrasted MPFB and MPFB+ participants, following the same procedure as had been used to
analyze the processing times. An ANOVA with Text and
Imagery Capacities as between-participant factors, and
Review as within-participant factor, revealed that reviewing times were overall longer for low than for high visuospatial imagers, 26.64 s (sd = 11.20) vs. 14.08 s (sd =
11.01), respectively, F(1, 21) = 12.30, p < .005. There was
no significant interaction between Imagery Capacities and
any of the other variables.
4.2.3. Distance comparisons
4.2.3.1. Frequency of correct responses. Fig. 2 shows the
mean frequencies of correct responses in the distance-comparison task for each set of items. An ANOVA was carried
out with Gender and Text as between-participant factors,
and Magnitude of Distance Differences, Test, and
Response Type as within-participant factors. Text had no
significant effect on the frequency of correct responses. A
significant effect of the magnitude of distance differences
was found, F(2, 56) = 25.52, p < .001. Tukey’s post hoc
tests showed that there were significantly fewer correct
responses to D1 than to D2 items (p < .001) and to D2 than
to D3 items (p < .001). The frequency also differed between
Test 1 and Test 2, 0.82 (sd = 0.14) vs. 0.91 (sd = 0.10),
respectively, F(1, 28) = 14.94, p < .001. Lastly, the frequency of correct Yes responses did not differ from that
of correct No responses.
The next step consisted of testing the impact of imagery
capacities. The ANOVA took into account Text and Imag-
205
ery Capacities as between-participant factors, and Magnitude of Distance Differences and Test as withinparticipant factors. Overall performance was lower for
MPFB than for MPFB+ participants, 0.83 (sd = 0.10)
vs. 0.89 (sd = 0.08), F(1, 21) = 3.61, p = .07. In addition,
MPFB participants appeared to be more obviously outstripped by MPFB+ participants for D1 than for D2 and
D3 items. The interaction between Imagery Capacities
and Magnitude of Distance Differences was significant,
F(2, 42) = 4.18, p < .05. Fig. 3 displays this interaction pattern. Inspection of the data revealed that performance was
significantly lower in MPFB than in MPFB+ participants
for the D1 items (Tukey, p < .001), but that the difference
was not significant for the D2 or D3 items.
4.2.3.2. Response times. The response times for each item
were recorded. Only the times of correct responses were
taken into account in the analyses reported below. Individual times which were above the mean time of the individual’s group by more than 2.5 sd in one of the two
comparison tests, for one of the subset of items (D1, D2,
D3) eliciting either Yes or No responses, were replaced
by that extreme value (m + 2.5 sd). Such cases only
accounted for 1.6% of the total number of individual
response times.
Fig. 4 shows the mean response times in the distancecomparison task for each set of items. The same steps were
followed for the analysis of the response times as for the
analysis of the frequency of correct responses. The first
ANOVA involved Gender and Text as between-participant
factors, and Magnitude of Distance Differences, Test, and
Response Type as within-participant factors. Responses
tended to be faster after processing of Text 1 than Text
2, 2840 ms (sd = 920) vs. 3121 ms (sd = 767), but the effect
remained below the level of significance. A significant effect
of the magnitude of distance differences was found,
1.00
0.93
Frequency of correct responses
0.87
0.80
0.80
0.60
0.40
0.20
0.00
D1
D2
D3
Magnitude of distance differences
Fig. 2. Mean frequency of correct responses as a function of the magnitude of distance differences (Experiment 2).
206
M. Denis / Acta Psychologica 127 (2008) 197–210
1.00
0.89
Frequency of correct responses
0.85
0.94
0.92
0.84
0.80
0.72
0.60
MPFBMPFB+
0.40
0.20
0.00
D1
D2
D3
Magnitude of distance differences
Fig. 3. Mean frequency of correct responses as a function of the magnitude of distance differences for MPFB
and MPFB+ participants (Experiment 2).
4000
Response time (ms)
3500
3419
2953
3000
2570
2500
2000
D1
D2
D3
Magnitude of distance differences
Fig. 4. Mean response times (ms) as a function of the magnitude of distance differences (Experiment 2).
F(2, 56) = 40.33, p < .001. Tukey’s post hoc tests showed
that response times were significantly shorter for D1 than
for D2 items, and for D2 than for D3 items (in both cases,
p < .001). Response times also differed for Test 1 and Test
2, 3248 ms (sd = 868) vs. 2714 ms (sd = 896), F(1, 28) =
33.20, p < .001. Lastly, times were systematically shorter
for Yes than for No responses, 2867 ms (sd = 895) vs.
3095 ms (sd = 889), F(1, 28) = 5.28, p < .05.
The next analysis tested the impact of imagery capacities. The ANOVA took into account Text and Imagery
Capacities as between-participant factors, and Magnitude
of Distance Differences, Test, and Response Type as
within-participant factors. Fig. 5 shows the mean response
times for each set of items and each group of participants.
Overall, response times were significantly longer for
MPFB than for MPFB+ participants, 3473 ms (sd =
798) vs. 2750 ms (sd = 856), F(1, 21) = 6.26, p < .05. There
was no significant interaction between Imagery Capacities
and Magnitude of Distance Differences. The same pattern
of results was obtained when the so-called predictable
items were withdrawn from the analysis.
4.3. Discussion
The first major finding of Experiment 2 was to provide a
further demonstration of the symbolic distance effect during
the processing of visual images, in the special case where
these images have been generated from a verbal description
M. Denis / Acta Psychologica 127 (2008) 197–210
4000
207
3890
3579
Response time (ms)
3500
3224
MPFBMPFB+
2951
3000
2636
2500
2391
2000
D1
D2
D3
Magnitude of distance differences
Fig. 5. Mean response times (ms) as a function of the magnitude of distance differences for MPFB
rather than being reconstructed from long-term memory (as
was the case in Paivio’s and Marschark’s pioneering studies). The results of Experiment 1 were confirmed in a new,
slightly different experimental context, including (a) longer
periods of time devoted to learning the configuration, (b)
variations in the text describing the configuration, and (c)
repeated distance-comparison tests. As well as confirming
the symbolic distance effect, Experiment 2 provided a second major confirmation, that of the effect of individual
imagery capacities on distance comparisons. More specifically, we obtained a further demonstration of the special
difficulty experienced by low visuo-spatial imagers in processing the most subtle distance differences.
The new feature of Experiment 2 was that it provided
information about the time required to perform mental
comparisons of distances. As the distance differences
increased, not only did the frequency of correct responses
increased, but response times also systematically decreased,
contributing to the signature of the symbolic distance effect
in our study (cf. Noordzij & Postma, 2005). Furthermore,
response times were shown to be affected by individual
imagery capacities. People with the highest imagery capacities had shorter response times than low imagers, which
suggested that they found it easier to access the metric
information contained in their mental images. Presumably,
their images contained more accurate information, which
increased the likelihood that they would produce correct
responses, and produce them in shorter times. However,
the chronometric measures did not corroborate any special
difficulty of low imagers in processing the smallest differences. Only the measures of performance revealed this difficulty (as was the case in Experiment 1), and chronometric
measures did not detect any such limitation of their capacities. Response times for the three levels of difficulty were
ranked similarly in the two groups, and low imagers systematically took longer to respond than high imagers.
and MPFB+ participants (Experiment 2).
Other effects of interest were also demonstrated by this
experiment. First, we carried out two distance-comparison
tests during the experimental session, something that does
not seem to have been done in previous experiments on
the symbolic distance effect. Not surprisingly, performance
increased and response times decreased from the first to the
second test, but the major points noted here were as follows: (a) the typical pattern of symbolic distance effect
was unaffected, a result which speaks in favor of the
robustness of the phenomenon, and (b) the effect was not
qualified by the level of individual imagery capacities. Second, we did not obtain any evidence that items requiring
No responses would elicit performance patterns that differed from those requiring Yes responses. No difference
emerged in terms of the frequency of correct responses.
However, a clear effect was found for response times. It
took longer to produce a negative response than a positive
one. This effect was consistent, and did not interact with
other variables of interest, notably with the magnitude of
distance differences or the participants’ imagery capacities.
The newest feature of the present study was to include
measures related to the acquisition of the representation
on which the symbolic distance effect would later be tested.
In particular, the objective was to establish whether imagery capacities, which are known to impact on distance comparisons, would also play a role earlier in the process. The
construction of visuo-spatial representations from the
descriptive texts has been amply documented in previous
research, which has established that people are able to create images with properties similar to those derived from
visual perception (Denis et al., 1995; Mellet et al., 2002;
Péruch et al., 2006). The descriptions used in Experiment
2 are the same as those used to demonstrate their impact
on the generation and scanning of images (Denis & Cocude, 1992). It was expected that greater difficulty in building coherent images would occur when reading these texts,
208
M. Denis / Acta Psychologica 127 (2008) 197–210
where information was presented in an unexpected (random) order (Denis & Cocude, 1997; Denis & Denhière,
1990). In the present experiment, during the processing of
the texts, a minor effect was detected, reflecting that the
random text did indeed take longer to process than the
well-structured one, but the effect remained below the level
of significance. Furthermore, high and low visuo-spatial
imagers did not display differing processing strategies, at
least in terms of reading times.3
The task where the impact of individual differences was
quite obvious was the mental revision of the representation
under construction. It always took longer to reconstruct
and review the image of a configuration that had not been
described in a regularly organized fashion (clockwise).
When the participants could no longer read the description, the cognitive cost of reconstructing the configuration
on the basis of a memorized poorly structured text became
evident. Furthermore, imagery capacities had a significant
impact, with high imagers being better able than their
counterparts to reinstate and review the image that they
had constructed based on a description.
Lastly, no significant effect of text structure was detected
in distance comparisons. The representation was constructed under more difficult conditions from the poorly
structured text, but once it had been achieved, the representation resulting from the processing of Text 2 was available
to inspection in the same way as that derived from Text 1.
The dominating factor was the individuals’ imagery capacities, the impact of which was the most clearly assessed during distance comparisons. Most of the differences between
high and low imagers did not result from the difficulties
experienced when constructing the representation, but
reflected the greater difficulty experienced by the latter
group to access the metric information available in their
mental images.
5. General discussion
People’s ability to construct images from verbal information is a well-established capacity. Not only can such
images be constructed and contain information, but, more
3
An original aspect of the study is its combined approach to the
processes of encoding and those involved in the comparison task. A
covariate approach would be sound if there were a definite, motivated
expectation that text-processing times would be predictive of comparison
times. Although this approach is legitimate in principle, the present study
was not carried out to test any such hypothesis. We do, of course, know
that imagery capacities have an impact on the time needed to process a
descriptive text, and so time can be taken to reflect the cost (in terms of
cognitive resources) of constructing a representation. On the other hand,
imagery capacities can also be expected to have an impact on the
comparison process, but probably by tapping into resources of a different
nature. The resources involved in translating language into a representation are presumably distinct from those involved in visually inspecting the
representation. In an attempt to assess the validity of this assumption, we
tentatively calculated the correlation between the two sets of response
times (text processing and distance comparisons) for the whole set of
participants. No correlation was found between the two time measures.
importantly for our discussion, they can also provide accurate metric information. This has been clearly shown with
the help of the image-scanning paradigm. A relevant additional piece of information is provided by the demonstration of the symbolic distance effect in the present study.
By examining distance-comparison performance in representations derived from the processing of a map or a
description, Denis and Zimmer (1992) had already shown
that a basic psychophysical phenomenon such as the symbolic distance effect occurs regardless of whether the origin
of the image is perceptual or verbal. In the present study,
using an approach similar to that of Noordzij and Postma
(2005), we took another step in this direction. By focusing
on images constructed from text descriptions, we confirmed
the presence of the symbolic distance effect, while also testing the hypothesis of the imaginal substrate of the comparison task through the impact of individual imagery
differences on this task, a novel piece of information in
the symbolic distance literature that pertains to distance
comparisons.
The existence of the symbolic distance effect is compatible with the idea of an analogical representational substrate, although it does not in itself provide a definite
demonstration that mental imagery is used in distance
comparisons. By contrasting high and low imagers’ performance, we have provided a more valid approach to this
question. In the two experiments reported here, we have
shown that when distance comparison is applied to a representation that has been constructed from verbal information, high imagers consistently perform better, and their
response times are always shorter than those of low imagers. This finding supports the claim that visual images
underlie the distance-comparison process. It is important
to note that the symbolic distance effect is demonstrated
by both groups of participants (see also Paivio, 1978),
which suggests that low imagers (who are not ‘‘non-imagers’’) do indeed attempt to perform comparisons on the
basis of an imaginal representation, but that they have
more limited resources for generating, manipulating, and
executing computations on these images than high imagers.
Another relevant finding is that the magnitude of the difference in performance of high and low imagers is greatest for
the most difficult items, those which involve subtle differences and therefore require more scrutiny. However, in
addition to the differences between the two groups of participants, it should be noted that based on a text, an analog
representation can be constructed that contains metric
information that was not explicit in the text used to construct it.4
The data collected here are highly compatible with an
imagery-based explanation for performance in size com4
These conclusions are in line with those of the studies documenting the
role of visuo-spatial working memory during the processing of more
complex spatial texts (cf. De Beni, Pazzaglia, Gyselinck, & Meneghetti,
2005; Deyzac, Logie, & Denis, 2006; Gyselinck, De Beni, Pazzaglia,
Meneghetti, & Mondoloni, 2007).
M. Denis / Acta Psychologica 127 (2008) 197–210
parisons (Dean et al., 2005). We must certainly consider the
possibility that the representations built by low imagers
result from impaired learning, which would explain why
they take longer to reconstruct the representation during
the distance-comparison task. If the representation is of
poor quality and not sharp enough, thus resulting in fuzzy
representations of distances, then no symbolic distance
effect should in fact be expected. The fact that both groups
displayed the symbolic distance effect indicates that the
representations they used were functionally equivalent.
The difference reflected by low imagers’ longer comparison
times is due to their difficulty in performing the relevant
computations on analog representations.
These conclusions are similar in some ways to those
obtained from scanning studies, but a difference between
the two sets of findings should be pointed out. In scanning,
high and low imagers differ in terms of their absolute
response times, and only high imagers actually display
the scanning effect. In contrast, in distance comparisons,
high and low imagers also differ from each other in terms
of absolute times, but both produce the same pattern of
results, i.e., all participants display the symbolic distance
effect. Subtle distance differences are harder for low imagers to detect (thus resulting in a specific interaction), but
the overall pattern is the same. This finding supports the
notion that low imagers use the same type of representational resources as high imagers, but that the difference
between them resides in the differing amounts of computational resources available to process these representations.
Acknowledgements
The author would like to thank Véronique Charlot, who
was very helpful in collecting the data of Experiment 1. The
author is also grateful to Ariane Tom for her assistance in
the statistical analysis of the data, to Grégoire Borst and
Monika Ghosh for their invaluable help in the preparation
of the manuscript, and to Albert Postma and three anonymous reviewers for providing valuable suggestions on an
earlier version of this article.
References
Afonso, A., Gaunet, F., & Denis, M. (2004). The mental comparison of
distances in a verbally described spatial layout: effects of visual
deprivation. Imagination, Cognition and Personality, 23, 173–182.
Barsalou, L. W. (1999). Perceptual symbol systems. Behavioral and Brain
Sciences, 22, 577–660.
Borst, G., Kosslyn, S. M., & Denis, M. (2006). Different cognitive
processes in two image-scanning paradigms. Memory and Cognition,
34, 475–490.
Chabanne, V., Péruch, P., Denis, M., & Thinus-Blanc, C. (2004). Mental
scanning of images constructed from visual experience or verbal
descriptions: the impact of survey versus route perspective. Imagination, Cognition and Personality, 23, 163–171.
Dean, G. M., Dewhurst, S. A., Morris, P. E., & Whittaker, A. (2005).
Selective interference with the use of visual images in the symbolic
distance paradigm. Journal of Experimental Psychology: Learning,
Memory, and Cognition, 31, 1043–1068.
209
De Beni, R., Pazzaglia, F., Gyselinck, V., & Meneghetti, C. (2005).
Visuospatial working memory and mental representation of spatial
descriptions. European Journal of Cognitive Psychology, 17, 77–95.
Denis, M., & Carfantan, M. (1985). People’s knowledge about images.
Cognition, 20, 49–60.
Denis, M., & Cocude, M. (1989). Scanning visual images generated from
verbal descriptions. European Journal of Cognitive Psychology, 1,
293–307.
Denis, M., & Cocude, M. (1992). Structural properties of visual images
constructed from poorly or well-structured verbal descriptions. Memory and Cognition, 20, 497–506.
Denis, M., & Cocude, M. (1997). On the metric properties of visual images
generated from verbal descriptions: evidence for the robustness of the
mental scanning effect. European Journal of Cognitive Psychology, 9,
353–379.
Denis, M., & Denhière, G. (1990). Comprehension and recall of spatial
descriptions. European Bulletin of Cognitive Psychology, 10, 115–143.
Denis, M., Gonçalves, M.-R., & Memmi, D. (1995). Mental scanning of
visual images generated from verbal descriptions: towards a model of
image accuracy. Neuropsychologia, 33, 1511–1530.
Denis, M., & Kosslyn, S. M. (1999). Scanning visual mental images: a
window on the mind. Current Psychology of Cognition, 18, 409–465.
Denis, M., & Zimmer, H. D. (1992). Analog properties of cognitive maps
constructed from verbal descriptions. Psychological Research, 54,
286–298.
Deyzac, E., Logie, R. H., & Denis, M. (2006). Visuospatial working
memory and the processing of spatial descriptions. British Journal of
Psychology, 97, 217–243.
Finke, R. A. (1989). Principles of mental imagery. Cambridge, MA: The
MIT Press.
Finke, R. A., & Pinker, S. (1982). Spontaneous imagery scanning in
mental extrapolation. Journal of Experimental Psychology: Learning,
Memory, and Cognition, 8, 142–147.
Guenther, R. K. (1998). Human cognition. Upper Saddle River, NJ:
Prentice-Hall.
Gyselinck, V., De Beni, R., Pazzaglia, F., Meneghetti, C., & Mondoloni,
A. (2007). Working memory components and imagery instructions in
the elaboration of a spatial mental model. Psychological Research, 71,
373–382.
Iachini, T., & Giusberti, F. (2004). Metric properties of spatial images
generated from locomotion: the effect of absolute size on mental
scanning. European Journal of Cognitive Psychology, 16, 573–596.
Kosslyn, S. M. (1973). Scanning visual images: some structural implications. Perception and Psychophysics, 14, 90–94.
Kosslyn, S. M. (1980). Image and mind. Cambridge, MA: Harvard
University Press.
Kosslyn, S. M. (1994). Image and brain: The resolution of the imagery
debate. Cambridge, MA: The MIT Press.
Kosslyn, S. M., Ball, T. M., & Reiser, B. J. (1978). Visual images preserve
metric spatial information: evidence from studies of image scanning.
Journal of Experimental Psychology: Human Perception and Performance, 4, 47–60.
Kosslyn, S. M., Margolis, J. A., Barrett, A. M., Goldknopf, E. J., & Daly,
P. F. (1990). Age differences in imagery abilities. Child Development,
61, 995–1010.
Kosslyn, S. M., Pinker, S., Smith, G. E., & Shwartz, S. P. (1979). On the
demystification of mental imagery. Behavioral and Brain Sciences, 2,
535–581.
Kosslyn, S. M., & Rosenberg, R. S. (2001). Psychology: The brain, the
person, the world. Boston: Allyn and Bacon.
Likert, R., & Quasha, W. H. (1941). Revised Minnesota Paper Form Board
(Series AA). New York: The Psychological Corporation.
Marschark, M. (1983). Semantic congruity in symbolic comparisons:
salience, expectancy, and associative priming. Memory and Cognition,
11, 192–199.
McGonigle, B., & Chalmers, M. (1984). The selective impact of question
form and input mode on the symbolic distance effect in children.
Journal of Experimental Child Psychology, 37, 525–554.
210
M. Denis / Acta Psychologica 127 (2008) 197–210
Mellet, E., Bricogne, S., Crivello, F., Mazoyer, B., Denis, M., & TzourioMazoyer, N. (2002). Neural basis of mental scanning of a topographic
representation built from a text. Cerebral Cortex, 12, 1322–1330.
Mellet, E., Bricogne, S., Tzourio-Mazoyer, N., Ghaëm, O., Petit, L., Zago,
L., et al. (2000). Neural correlates of topographic mental exploration:
the impact of route versus survey perspective learning. NeuroImage,
12, 588–600.
Moyer, R. S. (1973). Comparing objects in memory: evidence suggesting
an internal psychophysics. Perception and Psychophysics, 13, 180–184.
Moyer, R. S., & Bayer, R. H. (1976). Mental comparison and the symbolic
distance effect. Cognitive Psychology, 8, 228–246.
Noordzij, M. L., & Postma, A. (2005). Categorical and metric distance
information in mental representations derived from route and survey
descriptions. Psychological Research, 69, 221–232.
Noordzij, M. L., Zuidhoek, S., & Postma, A. (2006). The influence of
visual experience on the ability to form spatial mental models based on
route and survey descriptions. Cognition, 100, 321–342.
Paivio, A. (1971). Imagery and verbal processes. New York: Holt, Rinehart
and Winston.
Paivio, A. (1975). Perceptual comparisons through the mind’s eye.
Memory and Cognition, 3, 635–647.
Paivio, A. (1978). Comparisons of mental clocks. Journal of Experimental
Psychology: Human Perception and Performance, 4, 61–71.
Paivio, A. (1986). Mental representations: A dual coding approach. New
York: Oxford University Press.
Paivio, A. (1991). Images in mind: The evolution of a theory. New York:
Harvester Wheatsheaf.
Pazzaglia, F., & De Beni, R. (2001). Strategies of processing spatial
information in survey and landmark-centred individuals. European
Journal of Cognitive Psychology, 13, 493–508.
Péruch, P., Chabanne, V., Nesa, M.-P., Thinus-Blanc, C., & Denis, M.
(2006). Comparing distances in mental images constructed from visual
experience or verbal descriptions: the impact of survey versus route
perspective. Quarterly Journal of Experimental Psychology, 59,
1950–1967.
Pinker, S., Choate, P. A., & Finke, R. A. (1984). Mental extrapolation in
patterns constructed from memory. Memory and Cognition, 12,
207–218.
Pinker, S., & Kosslyn, S. M. (1978). The representation and manipulation
of three-dimensional space in mental images. Journal of Mental
Imagery, 2, 69–83.
Pylyshyn, Z. W. (1973). What the mind’s eye tells the mind’s brain: a
critique of mental imagery. Psychological Bulletin, 80, 1–24.
Pylyshyn, Z. W. (2002). Mental imagery: in search of a theory. Behavioral
and Brain Sciences, 24, 157–237.
Reed, S. K., Hock, H. S., & Lockhead, G. R. (1983). Tacit knowledge and
the effect of pattern configuration on mental scanning. Memory and
Cognition, 11, 137–143.
Richardson, J. T. E. (1999). Imagery. Hove, UK: Psychology Press.
Richman, C. L., Mitchell, D. B., & Reznick, J. S. (1979). Mental travel:
some reservations. Journal of Experimental Psychology: Human
Perception and Performance, 5, 13–18.
Shepard, R. N., & Cooper, L. A. (1982). Mental images and their
transformations. Cambridge, MA: The MIT Press.
Trojano, L., Grossi, D., Linden, D. E. J., Formisano, E., Goebel, R.,
Cirillo, S., et al. (2002). Coordinate and categorical judgements in
spatial imagery: an fMRI study. Neuropsychologia, 40, 1666–1674.
Trojano, L., Grossi, D., Linden, D. E. J., Formisano, E., Hacker, H.,
Zanella, F. E., et al. (2000). Matching two imagined clocks: the
functional anatomy of spatial analysis in the absence of visual
stimulation. Cerebral Cortex, 10, 473–481.
Trojano, L., Linden, D. E. J., Formisano, E., Goebel, R., Sack, A. T., &
Di Salle, F. (2004). What clocks tell us about the neural correlates of
spatial imagery. European Journal of Cognitive Psychology, 16,
653–672.
Zwaan, R. A., & Radvansky, G. A. (1998). Situation models in language
comprehension and memory. Psychological Bulletin, 123, 162–185.
Zwaan, R. A., & Taylor, L. J. (2006). Seeing, acting, understanding: motor
resonance in language comprehension. Journal of Experimental Psychology: General, 135, 1–11.